seriously overestimate the reaction rate, and, therefore, the production rate of the chemical species which enters into the next reaction in the photochemical chain. On the other hand, the enhancement of the reaction rate for two materials emanating from a common source and, therefore, occupying the same volume of the atmosphere, during the initial period of incomplete mixing, may also represent a significant de- parture from conventional simulation models.

It is hoped that this brief and necessarily incomplete dis- cussion will serve to demonstrate the importance of turbulent fluctuations of concentrations in atmospheric chemical re- actions. consideration of these effects in refining simulation models of these reactions appears to be important.

Nomenclature Ci = mass fraction of ith chemical species, ppm Dt = molecular diffusion coefficient for ith chemical

K = chemical reaction rate constant, cm3/sec-mol K,, K p = chemical reaction rate constants, l/ppm-sec Mi = molecular weight of ith chemical species, g/mol N = nondimensional ratio of characteristic times q 2 = u‘2 + u’2 + w’2, cm2/sec2 R,,@ = ordinary second-order correlation coefficient t = time, sec u’ , u ’ , w‘ = orthogonal components of turbulent fluid

y t = length along ith direction of a Cartesian coordinate

species, cm 2/sec

- - -

motion, cm/sec

system, cm

[ 1, - = averaged quantity ’ = departure from the average of the primed quantity

GREEK LETTERS

a, p, y, 6 = chemical species X = dissipation scale length, cm A = macroscale of atmospheric turbulence, cm po = dynamic viscosity for air, g/cm-sec po = fluid density, g/cm3

Literature Cited

Csanady, G. T., Murthy, C. R., J. Phys. Oceanogr., 1, 1, 17-24 (1971).

Donaldson, C: duP., J . AIAA, 7, 2, 271-8 (1969). Nickola, P. W., Ramsdell, J. V., Jr., Ludwick, J. D., “De-

tailed Time Histories of Concentrations Resulting from Puff and Short-Period Releases of an Inert Radioactive Gas: A Volume of Atmospheric Diffusion Data,” BNWL- 1272 uc-53 (available from Clearinghouse, NBS, U.S. Dept. of Commerce), 1970.

O’Brien, E. E., Phys. Fluids, 14, 7, 1326-9 (1971). Singer, I. A., Kazuhiko, I., del Campo, R. G., J . Air Pollut.

Contr. Ass., 13, 1, 40-2 (1963). Worley, F. L., Jr., “Report on Mathematical Modeling of

Photochemical Smog,” Proceedings of the Second Meeting of the Expert Panel on Modeling, No. 5, NATO/CCMS Pilot Project on Air Pollution, Paris, July 26-7, 1971.

Receiced for review Noaember 26, 1971. Accepted May 11, 1972.

Hydrogen Peroxide Formation from Formaldehyde Photooxidation and Its Presence in Urban Atmospheres

Joseph J. Bufalini,‘ Bruce W. Gay, Jr., and Kenneth L. Brubaker Environmental Protection Agency, Research Triangle Park, N.C. 2771 1

Hydrogen peroxide has been observed both in laboratory studies involving the photooxidation of formaldehyde and in urban atmospheres. In the laboratory studies, the H202 yield was nonlinear with formaldehyde consumption and decreased in the presence of oxides of nitrogen. The urban atmosphere at Hoboken, N.J., contained 4 pphm of H 2 0 2 under moderate smog conditions. Under severe smog conditions, the River- side, Calif. atmosphere contained 18 pphm of H202.

n the study of Purcell and Cohen (1967), hydrogen perox- I ide was found as a photooxidation product of formalde- hyde. This observation contrasts with the earlier work of Carruthers and Norrish (1936) as well as that of Horner and Style (1954) who observed no H202. The differences in results obtained by these investigators may arise from the concentra- tions employed in the laboratory irradiations. Purcell and Cohen worked with very low concentrations, ranging from 1 to 30 ppm (v/v), whereas the earlier workers used millimeter

To whom correspondence should be addressed.

816 Environmental Science & Technology

partial pressures of formaldehyde. This difference suggests that the mechanism for the photooxidation of formaldehyde may be significantly altered at low concentrations.

The purpose of this investigation was to extend the work of Purcell and Cohen by observing the H202 and all of the carbon fragments resulting from the photooxidation of formaldehyde at two different wavelength regions (maxima near 3100 and 3600& respectively) and in the presence and absence of oxides of nitrogen. Also, since our laboratory studies have detected H202 in irradiated systems containing oxides of nitrogen and such hydrocarbons as 1,3,5trimethylbenzene, ethylene, pro- pylene, 1-butene, as well as irradiated auto exhaust, we con- cluded that H202 may also be present in polluted urban at- mospheres.

Experimental Formaldehyde in air was irradiated either in a 72-1. boro-

silicate flask or in 160-1. plastic bags made from fluorinated ethylene-propylene copolymer (Du Pont FEP Teflon). The bags were constructed of Tedlar plastic in the Hoboken, N.J., study. The irradiations were carried out in a thermostatted irradiation chamber described elsewhere (Altshuller and Cohen, 1964) at a temperature of 23’ i 1’C and at ambient pressure. The irradiation chamber was fitted with either 24

value was all that was needed for our studies, but the absolute light intensity (photons cm-*sec-’) can be obtained by using the absorption cross section for NO2 and the quantum yield, The first-order dissociation constants kd for the chamber were 0.32 min-1 and 0.14 min-I for 24 blacklights and 28 sunlamps, respectively.

During the summer of 1970, the Environmental Protection Agency had one of its mobile instrument trailers stationed in Hoboken, N.J. Air was collected in large plastic Tedlar bags during the early morning hours of heavy traffic at the New Jersey entrance to the Lincoln Tunnel. These bags of polluted air were protected from sunlight and returned to the instru- ment trailer for irradiation and analyses. The collected air samples were irradiated by exposing the bags to sunlight on the roof of the trailer.

GE F-40-BLB blacklights with maximum intensity at 3600A or with 28 Westinghouse FS-40 sunlamps with maximum in- tensity at about 3100A. The irradiation container was filled by injecting either or both nitrogen dioxide and formaldehyde solution through a rubber septum onto the inner wall of a glass tube through which a gas, usually air, was passing, The glass was then heated gently to ensure complete evaporation of the reactants.

Analysis for formaldehyde was performed by the chromo- tropic acid method (Altshuller et al., 1961), nitrogen dioxide concentration was measured by the method of Saltzman (1954), and nitrate as nitrite by reduction with hydrazine in a hot alkaline solution. The nitrate (as nitrite) was then diazo- tized with sulfanilamide and N(1-naphthyl)-ethylenediamine dihydrochloride (Mullin and Riley, 1955).

Carbon monoxide and carbon dioxide were measured by gas chromatography with a catalytic convertor of the type origi- nally developed by Schwenk et al. (1961) to convert the CO to methane for easy detection with the flame ionization detector. The formic acid analyses were made with 1 -ethylquinaldinium iodide reagent (Sawicki et al., 1962). Hz02 was determined by using the specific method developed by Cohen and Purcell (1967). This method involves the reaction of titanium IV and 8-quinolinol with HzOz to form a colored complex. Possible interferences from various atmospheric pollutants (SO,, 03, NO2, NO, and olefins) were investigated. Only SO, at high concentrations was found to give a negative interference of 0 . 7 z at a SO, concentration of 20 ppm. Other compounds such as peracetic acid, ethyl hydroperoxide, n-butyl hydro- peroxide, acetyl peroxide, and peroxyacetyl nitrate (PAN) were tested by Cohen and Purcell and found to give a very slight negative interference at high concentrations. Since no one has reported any of these compounds, aside from PAN, as being present in the atmosphere, one can assume that interferences are minimal. The concentrations of PAN in urban atmosphere are usually quite low (-4 pphm). Thus, no PAN interference with the HzOz determinations is possible.

Singlet oxygen (0, ’Ag) was not tested as a possible inter- ference for H202 determinations in the atmosphere, but it is doubtful that 0, ‘Ag is present at sufficiently high concentra- tions to make its interference important.

The formaldehyde solution was Matheson, Coleman and Bell reagent grade 36-38Z in water with 10-15Z methanol added as a preservative. This solution was used without further purification. Baker dry air and prepurified nitrogen and nitrogen dioxide (Matheson, 99.5% pure) were used.

The stability of HZO2 in FEP Teflon bags was investigated by injecting microliter quantities of a standardized 22.9 % HzOz solution into a metered airstream while filling the bag. The theoretical concentration of peroxide was calculated from the amount of liquid injected and the volume of air used. The actual concentration in the bag was determined by both the titanium IV reagent and the catalyzed 1% potassium iodide method (Cohen et al., 1967).

Ozone at the New Jersey site was monitored with a Regener- type instrument (Regener, 1960) while total oxidant was measured with the neutral KI procedure. Laboratory measure- ments employed the chemiluminescence from the ethylene- ozone reaction (Nederbraght, 1965) and the Beckman Mark IV uv photometer. Total oxidant readings in California were obtained on a Mast ozone instrument. H2O2 gives a 7 z slow response with the Mast (Bufalini, 1968).

Light intensity of the irradiation chamber was determined by measuring the initial rate of disappearance of NO, in an inert atmosphere, as described by Tuesday (1961). This kd

Results During the early stages of this work, we were uncertain

about the suitability of formalin solution because of the presence of methanol. We compared the formaldehyde con- sumption and HzOz production of formalin and paraformalde- hyde. The results showed that the presence of methanol had very little, if any, effect on the photooxidation of formalde- hyde. In further tests, methanol was irradiated with NO,; these data are shown in Table I. Apparently some methanol reacts, since CO and formsldehyde are observed as products. The yield is small, however. Also, since the concentration of methanol present during the formaldehyde irradiations is three times lower than that shown in Table I, we concluded that methanol interference is minimal. We were not able to measure the methanol concentration directly. The retention time for the methanol on the gas chromatograph was ex- tremely dependent on the relative humidity of the air employed for the irradiations. Since FEP Teflon is extremely permeable to water vapor, we found that the sensitivity of the gas chro- matograph to methanol increased with increasing time. This phenomenon was apparently a result of column conditioning with water vapor.

The stability of HzOz in FEP Teflon bags was also investi- gated, The concentrations observed experimentally were lower than the calculated theoretical concentration. This discrep- ancy is probably owing to the destruction of H z 0 2 in the metal needle of the syringe and the destruction of HzO2 on the bag walls. An initial increase in HzO2 was observed. This phenome- non was reproduced in replicate runs and was probably due to desorption of H z 0 2 from the walls. In filling the bag ini- tially, the surface-to-volume ratio was very large and some H 2 0 2 apparently condensed on the walls. As the bag filled, the surface-to-volume ratio decreased, and HZOZ came off the walls and into the gas phase. The decomposition was linear at a rate of ppm/min at 2 ppm H 2 0 2 level. The data pre- sented in the figures and tables were not corrected for decom- position.

Data from the photooxidation of formaldehyde in the pres- ence and absence of NOz with sunlight fluorescent lights

Table I. Photooxidation of Methanol with Nitrogen Dioxide and Blacklights

Irradn Products formed, ppm MeOH, NO?, PPm ppm time, min co CHzO 12.0 1 . 3 60 0.253 0 12 .3 0 .27 120 0.153 0.280 1 2 . 8 0.20 120 0.109 0,242

Volume 6, Number 9, September 1972 817

IlWPlDlAflON TIME. minuds

Figure 1. Formaldehyde irradiated with and without NO? in the presence of sunlight

13 1 3

- 1 2

- 1.1

- 1.0 a

8 HCHO -

0.5 0" 0.4

0.3

0.1

0

8 3 = 2

O M 60 90 Im iRRADlATlON TIME. min.

Figure 2. Formaldehyde irradiated with blacklights and NO2

Table 11. Formaldehyde (12.5 ppm) in Air Irradiated with Sunlamps

Irradn time, (HCHO), (CO) WzOd min (HCH0)i (HCHO), (HCHO),

30 0.633 0.732 0.052 60 0.400 0.724 0.115

120 0,224 0.796 0.127 180 0.156 0.836 0,140 240 0.106 0.833 0.090

lRAAOlATlON TIME m m

Figure 3. Formaldehyde irradiated with blacklights

Table 111. Formaldehyde (12 ppm) Irradiated with 1.2 ppm NOs in Air and Sunlamps

30 0.503 0.771 0.073 0.133 . . . 60 0.336 0.824 0.0867 0.080 0.68

120 0.190 0.896 0.0871 0.039 0 .93 180 0.118 0.881 0.0713 0.030 0.81

Table IV. Formaldehyde (12.9 ppm) in Air Irradiated with Blacklights.

60 0.861 0.717 0 90 0.813 0.814 0.018

120 0.759 0.861 0.035 180 0,669 0.883 0.021

a Ozone <0.1 ppm after 180 min. This was not observed with a clean Teflon chamber.

Table V. Formaldehyde (12.7 ppm) Irradiated with 1.01 ppm of NO2 in Air and Blacklights

Irradn. time, ___ ___ ~ __ min (HCHO), (HCHO), (HCHO), (Nor), 0 3 , ppm

30 0.814 0.819 0.050 0.152 0.27 60 0.714 0.864 0.020 0.064 0.75

120 0.527 0.825 0.028 0.089 1.19

(HCHO) (CO) (HzOz) ( N o d

(wavelength maximum at 3100A) are shown in Figure 1. Re- sults from the photooxidation of formaldehyde again in the presence and absence of NO2 but wit! blacklight fluorescent lights (wavelength maximum at 3600A) are shown in Figures 2 and 3. The reduced data are shown in Tables 11-V for the various conditions.

Gas chromatographic analyses indicated CO as the major product. As shown in the tables, this product accounted for as much as 90% of the total carbon. Limited data from two experiments with a closed glass system indicated that between 12 and 2 0 z of the carbon appears as COS. The remaining carbon appeared as CO.

Various techniques were employed to test whether any oxi- dant other than Hz02 was formed in the photooxidation of formaldehyde. These tests were performed by the titanium-8-

818 Environmental Science & Technology

quinolinol method for H202 with neutral potassium iodide re- agent, the molybdate-catalyzed potassium iodide reagent, Mast instrument equipped with an olefin titration system, and a Beckman Mark IV uv photometer. After obtaining some con- tradictory results, we found that formaldehyde interfered with the molybdate-catalyzed potassium iodide reagent giving more oxidant than is actually present. No other oxidants, such as performic or peroxyformyl nitrate, were observed as products. Some ozone (-0.1 ppm) was observed in the photooxidation of formaldehyde in the presence of blacklights and in the absence of oxides of nitrogen. However, this phenomenon was dependent on the history of the Teflon container. A con- tainer that had not been previously employed in irradiations involving oxides of nitrogen did not produce ozone. This sug- gests that very small concentrations of NO, (ppb range) can

be responsible for significant concentrations of ozone. The H 2 0 2 yield was lower when formaldehyde was irradiated with sunlamps in the presence of oxides of nitrogen.

Analyses for formic acid were performed on several experi- ments under the various conditions of these studies. The formic acid data were inconclusive, probably because of the difficulty of the analytical method. In one experiment, no formic acid was observed when formaldehyde was reacted with NO, and sunlight for 30 and 60 min. At 120 min, when 9.69 ppm formaldehyde had reacted, 0.73 ppm formic acid was observed. At 180 min, when 10.16 ppm formaldehyde had reacted, 0.64 ppm formic acid was observed.

Since most if not all of the carbon from formaldehyde can be accounted for, it was necessary to determine the fate of NO,. We (Gay and Bufalini, 1971) showed recently that when hydrocarbons are photooxidized in the presence of oxides of nitrogen, much of the nitrogen oxide is oxidized to nitric acid on the walls of the reaction vessel. Since Teflon bags could not be used for this study, we repeated some of the work with a large borosilicate flask. Two experiments were performed with the large flask for this purpose. Formaldehyde (6 ppm) was irradiated for 6 hr with 2.3 ppm of NO,. After the irradiation, 2.25 ppm of nitrogen was recovered; 1.79 ppm appeared as nitrate on the walls; and 0.46 ppm appeared as nitrate in the gas phase. The alkaline hydrolysis product of NOs at low ppm concentrations is a nitrite. This experiment was repeated with 6 ppm formaldehyde and 1.9 ppm NOn. After 17 hr irradia- tion, no nitrate or nitrite were observed in the gas phase but 0.18 ppm of nitrite and 1.8 ppm of nitrate were observed on the walls of the reaction vessel.

The results from irradiating a bag of air collected at Hobo- ken are shown in Figure 4. The bag contained 6.0 ppm of CO, 4.1 ppm of methane, and 1.9 ppmC of nonmethane hydro- carbons. The total oxides of nitrogen were 36 pphm with 29.5 pphm as nitric oxide. After 1 hr exposure to sunlight, the NO, maximum was reached, and after 2 hr all of the nitric oxide and most of the NO, had disappeared. Total oxidant maximum and Hz02 maximum were observed after 33/4 hr exposure.

The atmosphere in the vicinity of the mobile instrument trailer in Hoboken, N.J., contained concentrations up to 4 pphm of H202. This concentration was measured on a day of high solar radiation and moderate smog formation. On days when solar radiation was low owing to cloud cover, no H202 was observed.

In August 1970, the urban atmosphere at Riverside, Calif., was sampled for H z 0 2 during days of photochemical smog

IRR4DIITION TIME, hOUI

Figure 4. Lincoln Tunnel air sample irradiated with natural sunlight

formation. On August 6, during a severe smog episode, con- centrations of oxidant as high as 0.65 ppm were observed with the Mast ozone meter (Figure 5). Hs02 reached a maximum of 18 pphm during the episode. This maximum appeared at the same time the total oxidant maximized.

The time at which the oxidant maximum for Riverside oc- curred was later in the afternoon than is usually observed in downtown Los Angeles. This difference can be explained by meteorological factors relating to the transport of air masses from west to east. Movements of polluted air mass and changes in photochemical smog formation are reflected in the observed changes in HzOz concentrations. Figure 6 shows the HzOz concentration as a function of time of day for various days at the Riverside site. On August 7, moderate to heavy smog buildup was observed for the air mass which moved eastwardly over Riverside. Maximum total oxidant observed was about 0.3 ppm between 2:OO and 3:OO P.M. A maximum concentration of 6 pphm of HzOz was observed. The polluted air mass which formed west of Riverside on August 11 never reached the sampling site owing to a change in wind direction. Less photochemical smog on August 10 was exhibited by increased visibility and lower oxidant readings. The maximum oxidant was observed at about 3:OO P.M. when 1 pphm of HzOz was measured.

Discussion The formation of HzOz can be explained by a sequence of

reactions involving either the photolysis of formaldehyde, nitrous acid, or nitrogen dioxide. The photolysis of formalde- hyde has been reported (Calvert and Pitts, 1966) in terms of two primary processes:

Fu i--. m-, TlYE OF DAY

Figure 5. Measured oxidant at Riverside, Calif., August 6, 1970

1 1 I I I I I I I I

TIME OF DAY

Figure 6. H2O2 concentrations at Riverside, Calif., in August 1970

819 Volume 6, Number 9, SeDtember 1972

HCHO + hv + H2 + CO

The importance of both processes a t different wavelengths has been discussed recently by McQuigg and Calvert (1969) and by DeGraff and Calvert (1967). When CO, oxides of nitrogen, water vapor, and high partial pressures of oxygen are present, the following sequence of reactions can also occur :

NO2 + hv + N O + 0 M

0 + oz - 0 3

0 + HCHO + O H + HCO M

H + 0 2 + HOz

HCO + 02 + HC03

H0z + HCHO + HzOz + HCO

HCO3 + HCHO + HC03H + HCO

HC03 .--j CO + HOz

HCO3 + COz + O H wall

HOz --+ destruction

2H02 + HzOz + 0 2

HC03H + CO + HzO + ' / 2 0 2

HC03H + COz + Hz + '/zOz

H202 + HzO + '/zOz

OH + HCHO + HCO + HzO

NO + HOz + NO2 + OH

0 3 + NO --+ NO2 + 0 2

Nos + 0 3 + N o s + 0 2

NO3 + NO2 + Nz05

N205 + HzO + 2"03 2HNOz NO + NO2 + HzO

HNOz 2 OH + NO

OH + N O HNOz

O H + NO2 A H N 0 3

O H + C O - + C O z + H

From the sequence of reactions as written, it isn't obvious how formic acid can be produced in the absence of nitrogen oxides. In their presence, Reactions 28 and 29 can occur

HCO3 + N O + HCOz + NO2 (28)

HCOz + HCHO + HCOzH + HCO (29)

This mechanism shows that formaldehyde participates in HZOZ formation (Reaction 8). However, formaldehyde is not necessarily needed for the production of Hz02. If a hydro- carbon containing a weak C-H bond is replaced for form- aldehyde in Reaction 8 and any hydrocarbon in Reaction 5 , then as long as atomic hydrogen is formed (by Reaction 27 when CO is present), H202 can be produced. Since polluted urban atmospheres contain NO,, CO, HzO, and various hy- drocarbons, H 2 0 2 has a number of ways of being generated. However, the photolysis of formaldehyde is probably the pre-

820 Environmental Science & Technology

dominant source of HZOz in the atmosphere. In this work, as in the work of Purcell and Cohen (1967),

the molecular balance between peroxide formed and aldehyde consumed is clearly lacking. This is especially true for pro- longed irradiations, as shown in Figures 1 and 2. This dis- crepancy can be explained by Reactions 12 and 16 in the absence of nitrogen oxides. In the presence of nitrogen oxides, the amount of Hz02 produced is significantly lower. This sug- gests that Reaction 18 is important. It is doubtful that Re- action 9 occurs, since no oxidants other than Hz02 and ozone have been observed. It is also probable that peroxyformyl radical (Reaction 7) is not formed at all and that the HCO radical decomposes directly after combining with oxygen to form the products of Reactions 10 and 11. Evidence for this is found in the recent work of Dimitriades (1969) who was un- able to produce peroxyformyl nitrate.

In one respect, the data shown in the tables do not fit the reaction scheme shown. In Tables I1 and 111, the third col- umns show the CO/HCHO ratios for various irradiation times in the absence and presence of oxides of nitrogen. The ratios are higher in the presence of nitrogen oxides and with increas- ing irradiation time. Reaction 27 shows O H radicals reacting with CO. If this reaction were important, there should be a relative decrease in CO concentration after prolonged irradia- tion. The ratio should also be lower when oxides of nitrogen are present. Unless the rate constant for OH and HCHO is much greater than 2 x 10-13 cc molecule-' sec-l (Dixon- Lewis et al., 1966), which is the OH-CO rate constant, it would appear that O H radicals are unimportant in this system. How- ever, comparison of Figures l and 2 shows that the rate of disappearance of formaldehyde is 50 to 200 times greater than the rate that can be explained by just the atomic oxygen reac- tion (Herron and Penzhorn, 1969; Niki, 1966). Since the H?O, yield is low in the presence of NO?, and since little if any formic acid was formed and no peracid was observed, the types of free radicals operative in this system are clearly limited. Therefore, it must be concluded that OH radicals are the most important free radicals in this system and that the OH-HCHO rate constant is much greater than that for C O and OH. This conclusion is not in agreement with the work of Avramenko and Lorentzo (1949) but seems compatible with the observations of Herron and Penzhorn (1969). The former found a rate constant of 10-13 cc molecule-' sec-' for the OH-HCHO reaction while the latter group found a lower limit of 6 X 10-12 cc molecule-1 set-1. This observation is also compatible with the recent work of Morris and Niki (1971). These investigators observed a rate constant of 1.5 X lo-'' cc molecule-1 sec-1.

A possible explanation for the increase in the CO/HCHO ratio arises from the reaction of methanol. In Table I with 12 ppm of MeOH and 1.3 ppm of NOz, approximately 2% CO arose from the photooxidation in only 60 min. Since 4 ppm of methanol are present in all of the formaldehyde ex- periments, it is reasonable to assume that the amount of methanol reacting should increase with increasing irradiation time. The C O concentration should increase thus making the CO/HCHO larger.

Reactions 22 and 26 are included in the reaction scheme as a possible source of nitrate formation. However, since no ni- trate was observed in the gas phase, the reactions either occur on the walls of the reaction vessel or they are of little impor- tance.

The data shown in Tables I1 and IV indicate that more Hzoz is formed when formaldehyde is irradiated with sunlamps (A = 3100A) than when irradiated with blacklights (A =

3600A). This suggests that Reaction 1 is less important a t the longer wavelength, an observation that is compatible with the work of McQuigg and Calvert (1969) and earlier in- vestigators and incompatible with the observations of DeGraff and Calvert (1967). This observation is also in qualitative agreement with the recent work of Calvert et al. (1972).

Literature Cited Altshuller, A. P., Cohen, I. R., Int. J . Air Water Pollut., 8, 611

Altshuller, A. P., Miller, D. L., Sleva, S. F., Anal. Chem., 33,

Avramenko. L. I.. Lorentzo, R. V.. Dokl. Akad. Nauk. SSSR,

(1964).

621 (1961).

69,205 (1949).

Science, 175,751 (1972).

Bufalini, J . J., ENVIRON. SCI. TECHNOL., 2,703 (1968). Calvert, J. G., Kerr, J. A., Demerjian, K. L., McQuigg, R. D.,

Calvert. J. G.. Pitts. J. N.. Jr.. “Photochemistry,” Wilev, _ _ _ . New York, N.Y., ~’371,1966. ’

(1936). Carruthers, J. E., Norrish, R. G. W., J. Chem. Soc., 1036

Cohen, I. R., Purcell, T. C., Anal. Chem., 39,131 (1967). Cohen, I. R., Purcell, T. C., Altshuller, A. P., ENVIRON. SCI.

DeGraff, B. A., Calvert, J . G., J . Amer. Chem. Soc., 89, 2247

Dimitriades, B., Bartlesville Petroleum Research Center,

Dixon-Lewis, G., Wilson, W. E., Westenberg, A. A,, J. Chem.

TECHNOL., 1,247 (1967).

(1 967).

Bartlesville, Okla., private communication, 1969.

Phys., 44,2877 (1966).

Gay, B. W., Jr., Bufalini, J. J., ENVIRON. SCI. TECHNOL., 5, 422 (1971).

Herron, J. T., Penzhorn, R. D., J . Phys. Chem., 73,191 (1969). Horner, E. C. A,, Style, D. W. G., Trans. Faraday Soc., 50,

McQuigg, R. D., Calvert, J . G., J . Amer. Chem. Soc., 91, 1590

Morris, E. D., Jr., Niki, H. ,J . Phys. Chem.,75, 3640 (1971). Mullin, J. B., Riley, J. R., Anal. Chim. Acta., 12, 464 (1955). Nederbraght, G. W., Nature, 206,87 (1965). Niki, H.,J. Chem. Phys., 45,2330 (1966). Purcell, T. C., Cohen, I. R., ENVIRON. SCI. TECHNOL., 1, 845

Regener, V. H., J . Geophys. Res., 65,3975 (1960). Saltzman, B. E., Anal. Chem., 26,1949 (1954). Sawicki, E., Stanley, T. W., Pfaff, J . , Ferguson, J., “Analytical

Chemistry,” Proceedings Feigl Anniversary Symposium, Birmingham, Elsevier, Amsterdam, The Netherlands, pp

Schwenk, V.. Hachenberg, H., Forderreuther, M., Brennst.

1197 (1954).

(1 969).

(1967).

62-9 (1962). -

Chem., 42,295 (1961). Tuesdav. C. S.. “Chemical Reactions in the Lower and Umer

Atmosphere:” R. D. Cadle, Ed., Interscience, New YGrk, N.Y., pp 1-49,1961.

Receiaedfor reGiew October 20,1971. Accepted March 3, 1972. Presented in part at the Symposium on Chemical Reactions in Urban Atmospheres at the Research Laboratories, General Motor Corp., Warren, Mich., October 6 and 7, 1969, and the 161st National Meeting, ACS, Los Angeles, Calif., March 28- April 2, 1971.

Aerosol Filtration by Fibrous Filter Mats

Leonard A. Jonas,’ Carlye M. Lochboehler, and William S. Magee, Jr. Chemical Laboratory, Edgewood Arsenal, Md. 21010

w The physical properties and aerosol filtration characteristics of U S . Army Types 5, 6, 7, and 8 filter mats were studied. Filter bulk densities ranged from 0.197 t o 0.243 g/cm3, and the fractional void volumes from 0.845 to 0.883. Pressure drops of the filters were expressed as a function of the product of the mat thickness and the superficial linear velocity in a simple polynomial in conformity with Darcy’s law for fluid flow through porous media. Filtration by the filter mats of 0.3 p diam dioctylphthalate (DOP) aerosols was expressed both by the long-standing equation relating DOP penetration to the filter pressure drop a t any discrete flow velocity and by the more recent Dorman equations. The Dorman flow regimes were found consistent with the experimental data, and the parametric relationships confirmed the concept of a discrete velocity a t which maximum aerosol penetration of a filter occurred.

uch current interest is centered at the control and M reduction of air pollution. One of the means of accomplishing this is the removal of fine aerosol particles from the air by fibrous filter mats.

The published literature in the general field of aerosol filtra- tion is exceedingly voluminous, and the problem existing with

To whom correspondence should be addressed.

the new experimentalist is how to limit the references in order not to stagger the reader without slighting the important con- tributions of many workers in this field. Excellent background in aerosol filtration can be found in the work of First and Silverman (1953), Chen (1955), Green and Lane (1957), Fuks (1964), Werner and Clarenburg (1965), Pich (1965), Dorman (1966), and Torgeson (1968).

In addition to the generalized studies of aerosol filtration, a body of literature (“Handbook on Aerosols,” 1950; “Hand- book on Air Cleaning,” 1952) had accumulated on methods of generating aerosols, measuring their particle size distribution, the choice of solid or liquid aerosol, instrumentation for the detection of aerosol penetration of filters, and on the under- lying theories based on which the instrumental methods were developed.

Purpose The purpose of this study was to determine the penetration

characteristics of a 0.3-p-diam liquid dioctylphthalate (DOP)

aerosol through four types of U S . Army mask filter mats as a function of variables such as flow velocity, filter thickness, and pressure drop, and to develop those mathematical rela- tionships in accord with present theory which quantitatively expresses the experimental data.

Theory

Dorman (1960a, 1966) has developed a means of analyzing aerosol filtration data in terms of a semiempirical formulation

821 Volume 6, Number 9, September 1972